In the related art, a technique is known which detects positions and states of eyes based on a face image which is an image of a face of a person is captured. For example, JP-A-2001-022933 discloses a face image processing apparatus which accurately extracts nostrils and eyes at high speed using a two-dimensional template which shows characteristics of nostrils and eyes, without being influenced by attached information. The face image processing apparatus estimates an area where candidates of the eyes are present from a face image using the two-dimensional template, and extracts the candidates of the eyes by converting the candidates of the eyes in the area into data including a plurality of brightness values. Also, the face image processing apparatus determines the eyes and positions of the eyes by setting a two-dimensional center of gravity position of pixels of the candidates of the eyes to the position of the eyes in the extracted candidates.
In addition, JP-A-H06-255388 discloses a driving state detecting apparatus which reliably detects a driving state of a driver, even when brightness is changed in accordance with various traveling circumferences by processing a captured face image of the driver. The driving state detecting apparatus generates a target face template and a target eye template for the driver by regularizing concentration of the captured image of the driver and correlatively calculating the regulated image using a standard template in which areas of a face and eyes are specified in advance. The driving state detecting apparatus performs correlation calculation with respect to the regulated image using the target templates, and detects the driving state of the driver based on the change of a correlated value. Regardless brightness of vicinity, an abnormal state such as dozing off or inattention of the driver can be reliably detected by performing the regularization and the correlation calculation.
In addition, JP-A-2011-086051 discloses an eye position recognizing apparatus which is capable of flexibly corresponding to individual differences by reducing false recognition. The eye position recognizing apparatus acquires edge images by detecting edges in a face image, for example, calculates a degree of coincidence of a convex semicircular shape of which upper eyelids are simulated, with template and a predetermined image area in the edge image, and determines an image area where the degree of coincidence is equal to or more than a threshold, as a position of the candidates of the eyes. Then, edges of the positions of the candidates of two eyes, which are simultaneously moved to up and down, are considered as blinking of eyes, and the positions of the candidates of these eyes are determined as positions of the eyes.
In addition, JP-A-H09-276319 discloses an ophthalmic stereoscopic microscope apparatus which aims to qualitatively obtain a surface shape of eyes to be examined by observing stripe marks shown in the eyes to be examined using polarization light. In the ophthalmic stereoscopic microscope apparatus, if the eyes to be examined are observed which is irradiated with light of a light source through a polarization plate, or the like acting as a polarizer, various stripe patterns can be observed in accordance with states of corneas, and shapes of the corneas by the stripe pattern are determined.